Nonvolatile memory cell

The present invention is related to the following invention which is assigned to the assignee of the present invention:

A Regulated Programming Circuit for a Nonvvolatile Memory, having Ser. No. 07/666,964, and filed on Mar. 11, 1991;

The present invention relates, in general, to memory cells and, more particularly, to nonvolatile memory cells.

One particular type of nonvolatile memory cell is the Flash EEPROM (Electrically Erasable Programmable Read-Only Memory). Flash EEPROMs are a type of device which provides electrical erasing. The term "flash" refers to the ability to erase the memory cells simultaneously with electrical pulses. Typically, the erase operation erases the entire memory array (i.e. single transistors are not erased alone).

Known flash EEPROMs are typically programmed by using a cell structure which utilizes a programming voltage of approximately 10 to 15 volts connected to a control gate. These EEPROMs also use a drain voltage of approximately seven to nine volts. Under this operating condition, the conventional flash EEPROM implements a conventional hot carrier injection to raise the threshold voltage of all transistors being programmed.

One problem in the art of EEPROMs is that of over erasing the cell. If a cell is over erased, it will cause false results when aligned bits are read. For example, if a bit of word one is being read and the same bit of word two has been over erased, then current will flow through the transistor for the bit of word two resulting in a logic 0 reading regardless of the logic state of the bit of word one.

Two methods have been used to overcome this problem. The first is the use of an intelligent erase feature. Here, the system measures the total voltage of the array when programed with a logic 0 state (charged). A pulse erasing signal is then applied to the array and the the cells are checked again. This is repeated until none of the cells register a programed response following a pulsed erase signal. However, the problem here is that by the time the last cell is erased, the first cells may be entering an over erased state. This is compensated for by either making the voltage required to trigger the sense amp much higher, giving more room between a zero voltage level and the sense amp level, or requiring a much tighter control over processing so that there is little difference between those cells easily erased and those which require the most pulses. Neither of these options is particularly acceptable.

Another solution to this problem is the use of a series enhancement transistor. This transistor will act as a switch to block the current flow through a bit of a word that is not being accessed. While this helps solve the problem of obtaining erroneous results from over erased bits, it creates other problems of its own. For example, the cells must now be programed and erased using the drain, as opposed to programing with the drain and erasing with the source. Since the ideal conditions for programing and erasing are some what different, this means that a compromise must be reached which would not be ideal for either function.

An advantage of the present invention is that it permits programming through the drain and erasing through the source while preventing the over erasing problem. A nonvolatile memory cell embodying the present invention comprises a floating gate having two portions. The first portion overlies the channel region formed between source and drain. Part of the control gate overlies the first portion of the floating gate and part overlies a remaining portion of the channel region. A second portion of the floating gate extends from the first portion to a thin oxide tunnel area of the source. The cell further comprises an additional implanted region below the floating gate in the source tunnelling area as well as an additional diode implant beneath the drain region.

FIG. 1 is a prior art flash EEPROM cell with a floating gate disposed entirely between the control gate and the channel region;

FIG. 2 is a cross-sectional view of a portion of the flash EEPROM cell of FIG. 1;

FIG. 3 is a prior art flash EEPROM cell with a two level control gate have the floating gate disposed between the upper level and the channel region;

FIG. 4 is a cross-sectional view of a portion of the EEPROM cell of FIG. 3;

FIG. 5 is a schematic representation of the EEPROM cell of FIGS. 3 and 4;

FIG. 6 is a self-aligned EEPROM cell embodying the present invention;

FIGS. 7 and 8 are cross sectional views of the EEPROM cell of FIG. 6;

FIG. 9 is a non-self-aligned EEPROM cell embodying the present invention and

FIG. 10 is a cross sectional view of the EEPROM cell of FIG. 9.

Referring initially to FIGS. 1 and 2, a prior art EEPROM cell, generally designated 10, is illustrated. Flash EEPROM 10 consists of a source region 13 and a drain region 11 formed in a substrate 12 and defining a channel region 14 therebetween. A contact 17 is shown coupled to drain 11.

Overlying channel region 14 is a floating gate 16 which is separated from substrate 12 by a dielectric 18. Overlying floating gate 16 is a control gate 19 which is separated from gate 16 by a second dielectric 15.

Cell 10 is the type of cell which is subject to the over erase problems. Erasing is accomplished by electron tunneling. A voltage of approximately 12 volts is applied to source 13 while gate 19 is grounded. The electrons trapped in gate 16 flow toward the cluster at the portion of the floating gate overlying the source region and tunnel from the floating gate to source 13 due to the high electric field across oxide 18. However, this process is not self limiting. Erasing usually leaves gate 16 positively charged due to the tunnelling out of not only the electrons used to program cell 10, but also some of the built-in background electrons, possibly changing the transistor to a depletion-mode transistor.

To avoid over erasing, "intelligent" erasing can be used as described above. Generally, this involves a series of erase and verify steps.

Referring now to FIGS. 3 and 4, a second prior art memory cell, generally designated 20, is illustrated. Cell 20 consists of a source region 23 and a drain region 21 formed in a substrate 22 and defining a channel region 24 therebetween. A contact 27 is shown coupled to drain 21.

Overlying a first portion of channel 24 is a floating gate 26 which is separated from substrate 22 by a dielectric 28. Overlying floating gate 26 and a second portion of channel 24 is a control gate 29. Control gate 29 is separated from gate 26 and substrate 22 by a second dielectric layer 25.

The cell design of FIGS. 3 and 4 provides a device which is not subject to the errors which can be caused by over erasing the cell. As shown in the schematic representation of FIG. 5, cell 20 consists of a memory transistor 32 and an enhancement transistor 34. In this arrangement, memory transistor 32 may still be over erased. However, when cell 20 is not accessed, transistor 34 acts as a gate to prevent a current, which would provide a false reading, from passing through cell 20.

While cell 20 resolves the problem of over erase errors, it has the drawback of requiring that programing and erasing be conducted through drain 21. This is necessary since source 23 is not situated close enough to gate 26 to permit tunneling to occur.

In FIGS. 6-8, a flash EEPROM cell, generally designated 50, embodying the present invention is illustrated. Cell 50 consists primarily of a drain 51, source 53, floating gate 56, control gate 59, and a diode implant 65. As shown in FIG. 7, these elements are laid out primarily as in FIG. 4 (except for diode implant region 65) with gate 56 overlying a portion of channel 54 and is separated therefrom by dielectric layer 58. Gate 59 is then disposed over the remaining portion of channel 54 and over gate 56, being separated from the channel and gate by dielectric 55.

One difference between cell 50 and cell 10 is the extension of floating gate 56 over a thin oxide tunnel area 62 above source 53. A cross sectional view of this portion of cell 50, FIG. 8, shows that floating gate 56 extends over a portion of source 53. Source 53 being illustrated here as two portions, 53 and 53'. In one particular embodiment, source 53 would be an N + type while 53' would be an N type, with substrate 52 being a P - type material.

This permits the device to be erased using the source as opposed to the drain, as required in the device of FIG. 3. This also permits the drain and source to be configured to allow the optimum conditions for programing and erasing, respectively. This optimization in turn allows for faster programing and erase times.

In FIG. 7, a cross-sectional view of EEPROM 50 along line 7--7 of FIG. 6 is illustrated. FIG. 7 illustrates an additional implant (diode) region 65. In one preferred embodiment, implant 65 is a Boron diffused area forming a P type region within P - substrate 52. This implant region forms a zener diode at the interface with drain 51. This diode provides a drain voltage clamping mechanism to reduce the V PP sensitivity on drain disturb.

In FIGS. 9 and 10 a second embodiment of the present invention illustrating a non-self-aligned EEPROM cell is illustrated. As can be seen in FIG. 9, the edge of floating gate 56' is not aligned with the edge of control gate 59'. This would result in an unwanted transistor being formed at area 70 where control gate 59' directly overlies drain 51. This transistor is unwanted because it introduces undesirable resistance when programing.

To eliminate the extra transistor, the drain region is divided into two portions, 51 and 51'. In a specific embodiment, using a p type substrate, region 51 would be an n + type while region 51' would be an N - type. This results in the extra transistor being shunted and the circuit would function the same as cell 50, FIG. 6.

Therefore, an EEPROM is provided in which programming is achieved through hot-electron injection on the drain side and erasure is achieved through Fowler-Nordheim tunnelling on the source side. An implanted N region below the floating gate in the source tunnelling region area reduces the total erase current by eliminating the band-to-band tunnelling of electrons. The over erase problem is eliminated by the incorporation of a split gate on the source side which maintains the memory cell in enhancement mode even when the floating gate region is driven into depletion.

Thus, it will be apparent to one skilled in the art, upon reviewing the foregoing description, that there has been provided in accordance with the invention, a device that fully satisfies the aims and advantages.

While the invention has been described in conjunction with specific embodiments thereof, it is evident that many alterations, modifications, and variations will be apparent to those skilled in the art in light of the foregoing description. Accordingly, it is intended to embrace all such alterations, modifications, and variations in the appended claims.